CAJ | 학술논문

提出了一种基于视频图像序列的人体微小管状血管血液流速自动测量方法.在提取血管中心线时采用基于图像帧序列方差的种子点自动选取方法,设计了基于Hessian矩阵的血管中心线提取的迭代算法;提取血管中心线后,根据连续帧图像的中心线灰度统计曲线,计算出血液流速.结果表明:该方法测量速度快、精度高,与专家测量结果之间的相关系数为0.901 4.
제출료일충기우시빈도상서렬적인체미소관상혈관혈액류속자동측량방법.재제취혈관중심선시채용기우도상정서렬방차적충자점자동선취방법,설계료기우Hessian구진적혈관중심선제취적질대산법;제취혈관중심선후,근거련속정도상적중심선회도통계곡선,계산출혈액류속.결과표명:해방법측량속도쾌、정도고,여전가측량결과지간적상관계수위0.901 4.
An automatic measurement of microvessel blood velocity based on video image frames is proposed.In order to extract vessel centerline automatically, the method of seed point auto selection is proposed using the pixels variance between image frames, then the centerline detection iterative algorithm based on Hessian matrix is designed After the extraction of vessel centerline, the blood velocity distribution is calculated accurately with knowledge of the graylevel curve of centerline and time consumption between frames. Experiment result shows the proposed method is a rapid and precise detection one, which compared with expert detection result with 0.901 4 correlation coefficient, and could be a useful technique for human microvessel blood velocity measurement.